Historically, the Indus borderlands were among the least populated geographical spaces of South Asia. The British census of 1911 estimated the population of the entire area of what is now Pakistan as no more than nineteen million, a tenth of today's.Footnote ¹ Dis-aggregated population figures for the nineteenth century point at a still lower total. Around 1830, Charles Masson estimated the number of people within the territories of the Khan of Kalat (Baluchistan minus the southeastern corner of Las Bela) to be 450,000, about a twentieth of today's figure.Footnote ² Sindh or ‘Sindia Deserta,’ according to British government records, had a total population in 1855 of just over one million, not much more than one-fiftieth of today's.Footnote ³ Mountstuart Elphinstone reported a population of 4.3 million Afghans in the entire kingdom of Kabul in 1815, a little more than one-tenth of today's.Footnote ⁴ Kafiristan never had more than 60,000 inhabitants in the nineteenth century, perhaps half of today's estimated 112,000.Footnote ⁵ The western Panjab in 1881 had a population of about 6.5 million, whereas it constitutes 56% of Pakistan's total population of over 180 million today.Footnote ⁶ The valley of Kashmir had about 500,000 inhabitants in 1846, against about seven million today.Footnote ⁷

It can thus be concluded that the Indus borderlands at the time of British annexation — between 1820 and 1850 — had an aggregate population of no more than approximately twelve million, possibly less. We can perhaps speculate that the Indus borderlands had a total population of no more than a few million on the eve of the Mongol conquests.Footnote ⁸

Thinly inhabited, the Indus borderlands boasted few important historical towns and cities at any time in their history. Urbanisation was not the main driving force of Islamisation. Although urbanisation began in the Indus borderlands in approximately the fourth-to-third millennium bc, there are no pre-historic or ancient cities in the Indus borderlands today. In effect, the pattern of random disruption of cities by the shifting mud and water masses of major river systems which is typical of much of monsoon Asia has nowhere been better demonstrated than here.Footnote ⁹ The Indus borderlands were the site of the most precocious urban development anywhere in India, but they never became a durably and densely urbanised region like the Mediterranean. In the wake of continuing shifts in the course of the Indus and its Panjab tributaries (a river system that has been moving westward over the long term), old towns and cities continued to die and new ones arose. The cities of the Maurya, Indo-Greek, Kushana and subsequent ages suffered the same fate as Harappa and Mohenjodaro.Footnote ¹⁰ In the eighth century ad, the Arab conquerors found nothing here resembling the great walled cities of Khurasan (such as Balkh, umm al-bilād or ‘mother of cities’), and the cities they themselves built, on a barely identifiable substratum, were obliterated in their turn by earthquakes and floods, often hardly leaving a trace.Footnote ¹¹ The ruins of the once famed Arab seaport of Debal are a case in point — they are now a partly excavated site among desolate salt flats on a former mouth of the Indus, about thirty miles from the present coast. Virtually all cities of Sindh throughout history suffered a similar fate.Footnote ¹² It was the same with Multan, Kabul (reduced to village status by earthquakes in the fourteenth century) and Ghazna (destroyed by earthquakes, floods and landslides in the eleventh and fourteenth centuries) and other cities in the western Panjab and the Afghan territories.Footnote ¹³ As a result, already in medieval times the Indus borderlands boasted remarkably few important cities and towns that went back to an ancient past, and none became celebrated centres of culture over extended periods of time. What cities and towns the British conquerors found here in their turn were mostly Hindu-dominated bazaar and banking centres and these were generally small, of recent origin, and short-lived to boot.Footnote ¹⁴ There were no towns or cities that played or could have played a major role in the religious transformation — the conversion to Islam — of the largely rural and pastoral Indus borderlands.

This does not mean, on the other hand, that the Indus borderlands were a ‘barbarian’ frontier or marginal to civilisation. There was no fundamental civilisational gap between the Indus borderlands and the rest of the subcontinent. In other words, there were no pre-existing socio-cultural and religious differences that can explain the divergent trajectories of Islamic development in South Asia's borderlands on the Indus on the one hand and its Gangetic heartlands on the other. In Kafiristan, it is true, we find an ill-defined mixture of polytheism and animism, embedded in a tribal society without a written language and without much internal differentiation — in a geographically extremely remote and almost inaccessible mountainous niche of the Hindu Kush — that perpetuated itself over many centuries, not without change altogether but in virtual immunity from the traditions of the surrounding Hellenic and Indo-Iranian worlds.Footnote ¹⁵ But in the other agricultural and settled parts of the borderlands — in the valley of the Indus proper, and, above all, in Kashmir — much of the pre-Islamic heritage does not appear to have been essentially different from that of the Indian heartlands. Indian culture penetrated deeply into the more accessible areas on the major trade routes to the west of the Indus from early times, and the Kabul wilayat was historically the most Indianised of all.Footnote ¹⁶ Here the first stupas appeared under Ashoka. The early medieval Zunbil kings and their kinsmen the Turk Shahis of Kabul continued to proclaim themselves Buddhists but are also on record to have banned cow slaughter and beef consumption and to have worshiped the Shaivite god Zun. The connection of ‘greater Gandhara’ with Shiva is well established. Necropolises at Swat, Bamiyan, Murki Khel (in the valley of Jalalabad), and under the Safed Koh (where human bones were so abundant on the soil that walls were made of them), as well as the record of Gabar fire temples and pyrethrae or altars throughout what is now eastern and southern Afghanistan and elements in the art of these regions and linguistic evidence also show persistent links with Iran.

Like Kabul, Kashmir absorbed Indian culture from ancient times.Footnote ¹⁷ Buddhism arrived in Kashmir under the Kushanas, and by the seventh century members of the Karkota royal family and private persons who founded Buddhist stupas and viharas appear to have endowed brahmanical establishments and shrines of Shiva and Vishnu with equal zeal. With its considerable agricultural resources, the Kashmir dynasty was able to mobilise the military manpower of the ecologically marginal areas of the western Panjab, Afghanistan and the Tarim basin, and thereby extend its political control beyond the valley to the west and south. Radiating outward from Shrinagar, by the eighth century the zone of Kashmiri political and cultural hegemony then reached into the whole of the upper Indus valley and northern Panjab, as also into Gilgit, Baltistan and Ladakh — the northern Kohistans which later became largely Tibetan-spoken. Propped up by the Tang Chinese, Kashmir became an imperial power under Lalitaditya Muktapida and briefly conquered the brahmanical capital of Kanauj in 733, along with much of India's agricultural heartlands. In the wake of this epoch-defining digvijaya, brahmans became predominant in Kashmir as religious specialists and administrative officials.

Sindh was politically subordinate to Kashmir in its imperial heyday and developed a similarly composite religious culture of Indo-Iranian and pagan elements.Footnote ¹⁸ In its widest sense, ancient Sindh comprised an area on both sides of the Indus reaching much farther to the north and west than modern Sindh does today, including Multan and much of the western Panjab, and varying portions of the rocky uplands adjoining Baluchistan and of the sand hills of the Thar. Zoroastrianism flourished in Sindh, and from ancient times onwards the great temple of Multan was a major centre of the cult of the sun, owing its origin and increasing popularity to the influence of Zoroastrian priests coming from Iran, and attracting offerings across ‘the five Indies’ and thousands of pilgrims from all countries, while providing food, drink and medicines to the poor and sick. Iranian elements have always been pervasive in the folk art, dress, and religious iconography of the nomadic and semi-nomadic populations of Jats and Gujars in Sindh and the Thar Desert. But we also encounter the ancient remains of Buddhist buildings throughout Sindh. It was again Ashoka who brought Buddhism to Sindh. The Maurya emperor endowed monasteries and stupas in places where the sacred traces of the Buddha's presence were found — some of these were still largely intact when Xuanzang visited them almost a thousand years later.Footnote ¹⁹

Buddhism continued to flourish in its monastic form in Sindh, and there is evidence of Hinduism as well in both Sindh and the Panjab from the Gupta period onwards. Xuan Zong, in the third decade of the seventh century, provides some remarkably detailed statistical information about the distribution of hundreds of Buddhist monasteries, mostly belonging to the Sammatiya or ‘Personalist’ school of Hinayana Buddhism, and hundreds of Hindu temples in the area.Footnote ²⁰ Hinduism in Sindh, at the time of Xuanzang, represented the particular orientation of the Pāshupatis, a sect associated with Shiva in his aspect of the ‘Herdsman’. The worship of this form of Shiva may have considerable antiquity in Sindh — seals uncovered at Mohenjodaro reflect a proto-Pāshupati form of Shiva — but the Pāshupata sect as such did not become prominent until a century before the Arab conquest. Deltaic Sind then became a stronghold of this cult, but Shiva was also worshiped in the north of the province.Footnote ²¹ Regarding the Hinduism of Sindh, there is also the compelling evidence of a historical process of brahman immigration in the one or one-and-a-half centuries prior to the arrival of the Arab conquerors, as well as of the widespread influence of brahman priests and officials at that time.Footnote ²²

The pre-Islamic heritage of the Indus borderlands thus offers no obvious clues to their conversion to Islam. Can we look for such clues in the oral and written traditions of the Muslim populations now inhabiting these areas? As it turns out, this is an even less promising approach to the problem. Remarkably enough, the traditions current among the convert as well as immigrant Muslim populations of the Indus borderlands — but this is by no means unique to them — tend to entirely obliterate their ‘infidel’ past and routinely advance claims of descent from the Companions of the Prophet or the early Caliphs, or claims of conversion in the earliest centuries of Islamic expansion, or any number of pious lies regarding the rise of Islam amongst them. There are countless claims of this kind.Footnote ²³ And there have been numerous debates about such claims of early conversion that have become exceedingly rancorous and polemical. Few of them are supported by any evidence whatsoever, and most of them are demonstrably false.

The subject of pious lies, conversion to Islam has remained an ill-understood and difficult subject because of the paucity of the evidence relating to it. Conversion is a combined process of religious and social transformation that took place gradually and often imperceptibly slowly, over a number of centuries, and as such did not attract much attention from contemporary chroniclers. Even so, the subject has attracted considerable attention in modern scholarship on South Asia.Footnote ²⁴ At first (prior to the 1790s) it was mostly denied that any conversion had occurred in South Asia at all — this was the correlate of the view that Indian Muslims were members of an immigrant or immigrant-descended community. Subsequently, it was often argued that conversion was ‘forced,’ or due to material and political incentives, and also that conversion was motivated by a desire to escape from the debilitating oppression of the Indian caste system. In addition, there were various efforts to explain conversion as a safety mechanism for people who had violated caste rules, or as the result of intermarriage between Hindus and Muslims, as well as the result of missionary efforts by charismatic saints or Sufis.

The currently most widely accepted theory of mass conversion to Islam in South Asia relates it to processes of agricultural sedentarisation of pastoral and other mobile groups on the periphery of the subcontinent. This theory has a fairly long pedigree, and we find important elements of it, for example, in Mountstuart Elphinstone's classic An Account of the Kingdom of Caubul (1815). A variant of it can be found in Irfan Habib's work on Jat conversion to Sikhism in the Panjab.Footnote ²⁵ In its most elaborated form, however, it is above all associated with the work of Richard M. Eaton.Footnote ²⁶ “Much more in keeping with the geography and chronology of Muslim conversions in India”, writes Eaton, “would be an understanding of mass conversion as a process whereby preliterate peoples on the ecological and political frontier of an expanding agrarian society became absorbed into the religious ideology of that society”.Footnote ²⁷ According to Eaton, it was not caste Hindus who most readily converted to Islam, but non-agrarian pastoral and forest peoples whose contact with brahmanism and caste had been perfunctory at best and who were becoming integrated in a sedentary, agrarian society. In such frontier regions, Islam was more a ‘religion of the plough’ than a ‘religion of the sword’. Or, as Eaton elsewhere summarised his findings: “Recent research suggests that the growth of sedentary agriculture in lightly Hinduised regions of India will tell us more about conversion than will the movement of medieval armies”.Footnote ²⁸ Fundamental to this approach is the idea that the processes of sedentarisation were the same in the western Panjab [and Sindh] and on ‘the Bengal frontier’. What happened in the western Panjab and Sindh is comparable, in Eaton's view, with what happened somewhat later (mostly in the early modern period) in east Bengal when non-agrarian forest people were integrated into an expanding agrarian society that was emerging in the wake of an eastward shift of the Ganges river system and converted to Islam in the context of expanding Muslim rule in the area. In fact, Eaton derives most of his evidence for this general theory of conversion from east Bengal.

This sedentarisation or ‘religion of the plow’ theory has the merit that it relates religious transformation to broader social change and to geographical factors over the long term. With its emphasis on rural developments, it is also useful as a corrective to the idea that Islam is fundamentally a religion of the city. Nevertheless, as a general explanation for mass conversion to Islam it may to some extent be valid for east Bengal — where an entirely new peasant society was created in medieval times — but not for the Indus borderlands.

There are three reasons why the theory does not work for the Indus borderlands. The first reason is that the Indus borderlands are the site of the earliest and most precocious agricultural development in South Asia. Sedentarisation long pre-dates Islamisation in all parts of the Indus borderlands. The second reason is that, as we have seen, the pre-Islamic rural populations of the Indus valley and Kashmir were by no means ‘lightly Hinduised’ but deeply integrated in Hinduism and Buddhism. The third reason is that in the Indus borderlands there continued to be, at least until the dawn of the twentieth century, many groups of nomadic pastoralists (including many Jats) who converted to Islam and had a more than nominal commitment to Islam but who never sedentarised.Footnote ²⁹ The stereotype of the impious nomad, like that of the nomad implacably hostile to settled life, has no general validity.Footnote ³⁰

I argue, by contrast, that the medieval development of the Indus borderlands should be compared to that of Anatolia and parts of the Persian plateau, not to that of Bengal. In effect, the reasons that Turkey (but not Greece) and Pakistan (but not India) are Muslim countries today are broadly similar.

In both cases an existing religious culture unravelled under the impact of protracted nomadic conquest and migration. As is well known, the Turkish conquest of Anatolia lasted for almost four centuries, during which the once unified, stable Byzantine administration was destroyed and much of Anatolia was transformed into a scene of continuous military struggle.Footnote ³¹ Most importantly, the Turkish conquest and colonisation by Turkish nomads of Anatolia destroyed the Greek church as an effective social, economic, and religious institution. Since religion permeated practically every aspect of medieval society, the sharp decline of the church was a major factor in the disappearance of the Byzantine character of Anatolia. Moreover, Christian society, which had been subjected to such disruption and dislocation in the wake of the Turkish nomadic invasions, became isolated from Constantinople and was deprived of ecclesiastical leadership in the provinces. The Balkans and Greece, by contrast, although subjected to the Turkokratia for centuries, remained Christian. Most analysts agree that this had a lot to do with the fact that the Turkish conquest of the Balkans was not nearly as prolonged or repeated a process as it was in Anatolia, and that particularly the colonisation by Turkman nomads was not as heavy here. In the Balkans the conquest and colonisation were affected by a strongly centralised government that aimed to restore order as quickly as possible, not by multiple warlords with large nomadic retinues that were often at war with each other as much as with their Byzantine opponents. Moreover, Balkan Christian communities were not left without ecclesiastical institutions for long. After the conquest of Constantinople the Greek patriarchate was restored at the head of the Orthodox-Christian communities of the Balkans.

What happened in the Indus borderlands is broadly comparable to this development. When the Arabs arrived in the early eighth century they occupied Makran and parts of the Indus valley in one brief campaign, and dislocations remained relatively limited.Footnote ³² No tribal mass migration of Arab nomads followed upon the military conquest of Sindh (as it had in Iraq between 638 and 656). As a result there was an Arab military occupation of the towns, and some control of the countryside, but no extensive Islamisation or conversion to Islam of the native population. There is no evidence that such conversion occurred. To the contrary, an authoritative and detailed Persian geographical work of 982 ad confirms that the Baluchi nomads of the hot, pastoral desert between Kirman and the Indus were all Muslims (they had converted to the faith even before their arrival) but says in so many words that in Sindh only the Arab capital town of Mansura had “Muslim inhabitants”.Footnote ³³ The beginning of Turkish raids and conquests under the Ghaznavids resulted in more significant demographic dislocations, especially in the western Panjab — in these centuries we read of some conversion to Islam among some specific populations in the Indus borderlands, but it was still quite restricted, and in most places sporadic.Footnote ³⁴ The same can be said about the conquests of the Ghurids and their Turkish slave generals in the twelfth and thirteenth centuries.Footnote ³⁵ In short, when the Mongol nomadic hordes began to invade the Indus borderlands in the early thirteenth century, they came into an area of which large parts had already been under Muslim rule for half a millennium but where as yet no significant conversion to Islam had occurred.

Mongol expansion in the Indus borderlands between 1221 and 1398 was, however, extremely destructive.Footnote ³⁶ Large and sprawling hordes of warlike nomads, incessantly in search of pasture for their innumerable horses, and accompanied by sheep, flooded the Indus borderlands everywhere — Kashmir includedFootnote ³⁷ — and reduced much agricultural land to grassland, effectively turning the entire area into an extension of nomadic Central Asia. A scene of almost continuous warfare and devastation ensued, which lasted for almost two centuries, and decimated the already sparse populations. The only area of the Indus borderlands where the Mongols did not penetrate was Kafiristan (although Timur led a largely unsuccessful expedition into it). Everywhere else, the Indus borderlands were devastated. As a result, the Mongol invasions undermined and destroyed the very basis of the social order and completely de-legitimated and dismantled the existing religious infrastructure.

They thus proved to be a turning point in the historical development of the Indus borderlands. What is more, they set off a widespread and general migration and dispersal of tribal populations of nomadic Mongols, Turks, Afghans and Baluchis over wide areas which did not really subside until as late as the sixteenth century.Footnote ³⁸ It is in these conditions that the gradual conversion to Islam began to take place. But equally significant is the fact that the Mongols did not penetrate anywhere beyond the Indus borderlands, at least not normally and not for long. For ecological reasons the Mongol nomads could not sustain the occupation of other parts of the subcontinent. The subcontinent at large was unsuited for the kind of extensive pastoral nomadism practiced by the medieval nomads and therefore never suffered a nomadic conquest as such. And correspondingly conversion to Islam was much more sporadic.

In Sindh and the Indus borderlands generally the rise of Islam took place in an ‘age of anxiety’ which began with the prolonged turmoil and disruption that accompanied the Mongol invasions and ended with the subsequent re-ordering of the entire political order and recovery within the context of expanding Muslim states after the Mongols withdrew. Islam came to these regions through the institutionalised cult of pirs or saints (also known as ‘holy men’, and sometimes as Sufis) and the shrines built over their tombs.Footnote ³⁹ The fundamental raison d’être of these shrines was religious — it was through them and their rituals that Islam was made accessible to non-literate peasants and tribesmen. It was believed that saints enjoyed a closer relationship with God than common devotees, and that the saints’ spiritual power and their ability to intercede with God on the devotees’ behalf outlasted the saints’ mortal lifetime and adhered to their burial places, which became centres of pilgrimage and worship. Over time, saintly charisma became increasingly disassociated from personal piety. Unlike Christian saints (and post-eleventh century Catholic priests and bishops), most Muslim saints married and had children. Their spiritual power was hereditary and became distributed among all their offspring, with special provisions made for the few who fulfilled religious obligations and became guardians of tombs. Inherited saintliness came to legitimise acquired wealth, including land, and social position, with the result that many saints gained an almost regal power. As shrines grew wealthy, so did the proliferating saintly families who succeeded the original saint over many generations.

But the saints were also community builders in a medieval society that had been shattered. The shrines and holy lineages of Muslim saints brought stability and defined rights and entitlements by political arbitration — they evolved into a ‘hagiarchy’ of hereditary saints in a near-anarchic tribal environment.Footnote ⁴⁰ In arbitration disputes the saints could serve as ‘professional neutrals’ because their lineage connected them to the founder of the religion shared by both sides involved in the dispute but excluded them from identification with either. The innumerable smaller shrines that were built along roads, and on hill tops, served as demarcation points for pastoral tribes, or they separated the pastoral from the agricultural realm. The saints themselves would mediate disputes between pastoral and agricultural clans, or between competing pastoral clans. This was especially important in Sindh and the western Panjab, where, in the wake of the Mongol invasions, the social and economic structure changed radically as tribes increasingly settled on the irrigated plains of the Indus valley and more and more land was opened up through the use of the Persian wheel and later by digging artificial canals. Such profound demographic restructuring was accompanied by constant conflicts about land in which the saints played a critical role. Moreover, in these circumstances visiting and becoming clients or devotees of a shrine resulted in or formalised rights of passage along a route or grazing rights. The saints could also back up their arbitration by the use of force and often went to war with thousands of troops, interceded with rulers, and served as diplomatic envoys. Furthermore, the ‛urs or annual pilgrimages and festivals associated with the saints’ death anniversaries promoted peaceful contact between normally hostile or alien tribes and other population groups because violence was prohibited during the pilgrimage and the pilgrims travelled under the protection of the saints, while the tombs themselves were at all times exempted from tribal violence and were places of haram. And during the ‛urs there would be extensive trade in marketable goods, while sporting events brought the diverse tribes together within a religiously organised framework. Markets would be held in these conflict-free zones, and the saints thus controlled trade as well. A situation developed in which saints played not only a very important role in political conflict resolution and the control of markets while providing places of haram but also played a vital role in bridging sectarian divides and thus in religious conflict resolution.

On a religious level the saints were as much sought after for their miraculous and healing powers as for their knowledge of Islam. Visitors to their shrines would ask for a cure for a disease, sons, wealth, or forgiveness of sins. Just like Christian saints, the Muslim saints reputedly performed miracles. At the same time, shrines became centres of an entire medieval mental health industry, where drugs would be used and psychiatrically troubled people, especially women, could find therapeutic relief. At the annual ‛urs festivals people would indulge in ecstatic dancing, screaming prayers or demon-induced obscenities or otherwise find license to temporarily defy the normal rules of social behaviour in other ways. And Muslim shrines, like Christian shrines, abrogated and at the same time often perpetuated pagan cults and pantheistic nature worship. Numerous Muslim shrines or ziyarats in the Indus borderlands are found to have been erected on religious sites that were originally those of the former ‘idolatrous’ inhabitants. Digging up the soil in the precincts of shrines often brought to light large quantities of buried ‘idols’. To a large extent, the Muslim saints replaced the numerous gods of the Hindus, their shrines replaced the temples, and their festivals superseded the old Hindu festivals.

It was in these various ways that Muslim saints and their shrines became part of an entirely reconfigured religious landscape throughout the Indus borderlands and became instrumental in the gradual ‘conversion’ of their inhabitants to Islam. Historical chronicles begin to describe the rural and pastoral populations in Sindh, Afghanistan, Kashmir, and other parts of the Indus borderlands as Muslims around the mid-fourteenth century, and we can deduce from the fragmentary information these chronicles provide that conversion to Islam was by then well on its way. Such ‘conversion’ was, of course, never ‘complete’ anywhere. There were still numerous Zoroastrians in the Kabul wilayat in the seventeenth century, while at least a quarter of the population of Sindh remained Hindus even as late as the nineteenth century (there are still millions today), and most of the people of Kafiristan did not convert to Islam until the late nineteenth century. Saints and their shrines, however, were to become the central elements of the religious life everywhere in the Indus borderlands between the thirteenth and sixteenth centuries, and kept proliferating afterwards, marginalising almost everything else that we normally associate with the urban and literate traditions of Islam to a degree hardly found anywhere in the Muslim world. Saints and shrines became exceedingly numerous here. Early British travellers recorded that along the Indus the shrines were “almost innumerable” and that the people there “have abandoned essentially the religion of Islam, and have become votaries of a new worship, that of shrines”. Towns like Rohri, Sehwan, Nasarpur and Thatta claimed to have ‘sawa lakh’ or 125,000 saints buried in their graveyards, with Makli Hills near Thatta acquiring the reputation of being the largest necropolis in the Muslim world. Exaggerated as these claims may be, it is probably true that by the end of the eighteenth century it had become virtually impossible to travel more than a few miles almost anywhere in the Indus borderlands without coming across the shrine of one saint or another.

It was not until the nineteenth century that reformist movements began to oppose the culture of saints and shrines, their extravagant pageantry, and above all the claims that they were intermediaries between man and God. This happened for the first time in 1826, when Peshawar and the Yusufzai territories became the centre of a formidable Islamic challenge to Sikh dominance under Sayyid Ahmad Shah of Rai Bareli (1786–1831). Sayyid Ahmad Shah rallied the Afghans in a jihād or ‘holy war’ against the Sikhs, in an attempt to create an Islamic state with fundamentalist Wahhabite credentials. He was the first of a number of outsiders — of whom Osama bin Laden and Al-Qa'ida were the last — to migrate to the Indus borderlands from elsewhere and create jihādist movements that were, and still are, sharply at odds with the religious culture of the local saints and their shrines. Unlike what is often believed, they are not a throwback to medieval times but a modern phenomenon.